Founding AI & Backend Engineer Full Time NEW

Nightwave AI is a law enforcement AI startup building tools that help detectives solve crimes more effectively, efficiently, and accurately.

The goal is simple: give investigators the kind of intelligence capability that only the largest agencies in the world can afford today, and put it in the hands of every department. We are a small, fast-moving team working on a problem that genuinely matters, and we are looking to connect with talented, technically driven builders who want to do serious work on a serious product.

We are committed to advancing modern investigative methods while ensuring ethical and efficient processes.

Role Description

Nightwave’s product surface is where AI becomes real to detectives. But the engine underneath is where the hard problems live. This role owns that engine.

You will be one of the founding engineers at Nightwave, working directly with the founders. Your primary focus is the AI and systems core: the agent harness that powers our investigative reasoning, the data pipelines that ingest messy real-world evidence and external sources, the ontology that structures everything we know about a case, and the architecture that holds it all together. Your secondary focus is helping the team move fast across the stack when the situation calls for it.

For this to actually work in the real world, the underlying systems have to be fast, reliable, secure, and smart. That is what you will be building.

You will own how Nightwave actually thinks. That is a huge responsibility and we need someone who feels the weight of it.

What You’ll Actually Be Doing

• Designing and building the core agent harness: orchestration, subagents, tools, Skills, memory, planning, evals, context management, feedback loops, and guardrails for investigative agents that operate on real case data
• Building and maintaining the data pipelines that ingest, normalize, and structure evidence from dozens of both structured, unstructured, and external sources — PDFs, documents, media files, exports from forensic tools, agency record systems and existing evidence management systems, and public data + open source intelligence
• Architecting the Nightwave ontology; the semantic backbone that lets agents reason about people, places, events, evidence, leads, theories, and relationships within and across cases in complex, sophisticated ways based on the task at hand
• Designing backend systems, microservices, functions and tools, modules, and API layers that the frontend and agents both depend on (primarily Python with FastAPI, with flexibility where it makes sense)
• Integrating with external data sources and open source intelligence tooling; figuring out what’s useful, what’s reliable, and how to pull signal out of a lot of noise
• Building the evaluation and observability infrastructure that tells us whether the agent is actually working, where it fails, and how to make it better; being able to create the best update, compare to the old version through rigorous evals and realistic manual use on real difficult criminal cases, then actually update, then iterate and repeat
• Making architecture decisions that will define how Nightwave scales to hundreds of agencies ; performance, speed, reliability, cost, and security all matter
• Working inside a CJIS-aware environment and helping shape how we handle sensitive law enforcement data responsibly and compliantly; must be a US citizen to handle this sensitive data and work at Nightwave
• Shipping. Constantly. We need someone who builds, learns, and rebuilds; not someone who designs for six weeks before writing code

Who You Are

• A strong systems thinker with deep backend fundamentals. You understand distributed systems, data modeling, simple and clear code, API design, and the tradeoffs that matter when you’re handling real, messy, high-stakes data
• Experienced with modern AI and agent systems; LLM orchestration, tool use, Skills, memory, context management, subagents, continuous learning and improvement, post-training with RL and fine-tuning, retrieval, evaluation loops, guardrails. You have opinions and know about what works and what doesn’t in production
• Fluent in Python and comfortable in the FastAPI + async ecosystem. Bonus fluency in whatever else is useful: Go, Rust, TypeScript, SQL, infrastructure-as-code (React for frontend)
• Someone who has built with ontologies, graph databases, RAG architectures, complex data pipelines, and knowledge graphs; or who has the depth to learn them fast (most important attribute we are looking for is speed of learning)
• Experienced with cloud infrastructure (AWS, GCP, or Azure) and the practical tradeoffs of running real systems; cost, latency, reliability, observability, security
• Security-minded. You think about what could go wrong with sensitive data before anyone has to ask; CJIS-compliance and top-tier security standards are monumental for Nightwave and any product working in this industry
• Entrepreneurial and self-directed. At this stage, the person building it is the person deciding how it should work. You need to be comfortable making architecture calls and owning them
• Someone who has shipped real systems and knows the difference between a demo that works once and software that holds up under daily use by real operators
• Preferably based in or willing to be in the Columbus, Ohio area (hybrid, with real in-person time)

Bonus If You Have

• Experience building agent frameworks or harnesses from scratch — or deep work with existing ones like LangGraph, Deep Agents SDK or Claude Agent SDK, or custom orchestration layers
• Background in high-stakes operational environments: law enforcement, defense, intelligence, fraud, healthcare, or anywhere correctness and security are non-negotiable
• Experience with CJIS, FedRAMP, SOC 2, or other compliance frameworks
• Familiarity with OSINT tooling, forensic data formats, or investigative workflows
• Knowledge graph or ontology design experience; Palantir Foundry or Gotham, Neo4j, RDF, or custom graph architectures
• Experience with large-scale unstructured (and structured) data pipelines — audio, video, documents, structured exports
• Time at a similar mission-driven startup, Palantir, Anduril, Flock Safety, or similar

What We Offer

• Competitive salary plus meaningful early-stage equity, with flexibility on how we structure the package between cash and equity depending on what matters most to you
• Direct work with law enforcement; the systems you build will power real investigations and solve real crimes
• A founder who is technical and will be building alongside you
• The chance to define how the engine powering modern investigations actually works